3D Reservoir Computing with High Area Efficiency (5.12 TOPS/mm 2 ) Implemented by 3D Dynamic Memristor Array for Temporal Signal Processing

作者：Wenxuan Sun, Woyu Zhang, Jie Yu, Yi Li, Zeyu Guo, Jinru Lai, Danian Dong, Xu Zheng, Fei Wang, Shaoyang Fan, Xiaoxin Xu, Dashan Shang, Ming Liu · 发表于：2022 IEEE Symposium on VLSI Technology and Circuits (VLSI Technology and Circuits) · 年份：2022 · DOI：10.1109/vlsitechnologyandcir46769.2022.9830310 · 被引用次数：18 · 研究领域：Neural Networks and Reservoir Computing、Advanced Memory and Neural Computing、Neural dynamics and brain function

In this work, we realized a three-dimensional (3D) reservoir computing (RC) by utilizing the I-V nonlinearity and short-term memory of the dynamic memristor in 4-layer vertical array. The cycle-to-cycle variation of the dynamic reservoir is improved by parallel memristor configuration. The dimensionality of the reservoir space is increased by input strategy design. After the hardware-software co-optimization, the proposed 3D RC system exhibits high recognition accuracy (90%), low energy consumption (~0.78 pJ /operation), and high area efficiency (5.12 TOPS/mm2).